AN EXAMINATION OF THE ANTITRUST ISSUES POSED BY GOOGLE'S ACQUISITION OF ITA - American Antitrust Institute
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
AN EXAMINATION OF THE ANTITRUST ISSUES POSED BY
GOOGLE’S ACQUISITION OF ITA
Randy Stutz*
February 18, 2011
EXECUTIVE SUMMARY
The proposed merger between Google and ITA Software, Inc. (ITA) offers antitrust enforcers the
latest in a series of escalating challenges in which high technology markets seem to be straining the
boundaries of antitrust analysis. Apart from the very difficult analysis, however, this transaction
raises broader questions of competition policy concerning Google’s rapid growth as a one-of-a-kind
firm. The deal might even have implications for the future of airline ticket distribution. In this
white paper, the American Antitrust Institute (AAI) seeks to identify and explore both the narrow
and broad competition issues that are raised by a Google/ITA combination.
In the narrowest sense, acquiring ITA would put Google in the business of supplying a technology
input that powers downstream products in a vertical online search market. That is, Google would
own what many consider to be the premier technology that online travel agents, travel meta-search
websites, and airline websites license from ITA to afford Internet users the ability to search real-time
pricing and seat availability data in the course of shopping for airline tickets online. Neither Google
nor ITA currently competes in the provision of this data to Internet users by “online travel search”
firms, but together they seem to have such firms surrounded. Companies like Kayak, Orbitz and
Hotwire, as well as airlines, rely heavily on Google to tell consumers where they can get airline
pricing and availability data on the Internet, and they rely heavily on ITA to deliver the data itself.
The Antitrust Division of the Department of Justice (DOJ or the Division) must therefore evaluate
the potential competitive effects of a transaction that does not directly accomplish vertical
integration in any assumed relevant product market or obviously eliminate actual (as opposed to
potential) horizontal competition in any assumed relevant product market, but yet strategically
positions a likely monopolist of a nearby and intimately related market to enter the online travel
search market with significant competitive advantages. At bottom, the Division will simply have to
*Randy Stutz is the Director of Special Projects and a Senior Fellow of the American Antitrust Institute (AAI), a non-
profit research, education and advocacy organization devoted to advancing the role of competition in the economy,
protecting consumers, and sustaining the vitality of the antitrust laws. See http://www.antitrustinstitute.org. For
questions or comments on this white paper, please contact rstutz@antitrustinstitute.org or bfoer@antitrustinstitute.org.
Publication of this White Paper was approved by the AAI Board of Directors.
1
determine whether the effect of the transaction may be substantially to lessen competition. Under
the Clayton Act’s standard, which calls for the prevention of anticompetitive dangers in their
incipiency, this determination will demand sound predictions within a complicated space.
AAI believes that vertical merger standards as they have evolved over the last two decades, along
with the potential competition doctrine as it is incorporated in the DOJ’s Revised Horizontal Merger
Guidelines, are up to the task of steering a proper analysis. The Division should use these tools to
explore the risk that Google may acquire market power in the online travel search market or the
technology input market, along with the risk that Google’s control of ITA would lead to foreclosure
or other exclusionary effects, whether directly or indirectly. The Division should also consider
whether the transaction might have the effect of raising barriers to entry into the broader online
search market, which Google already dominates.
Looking beyond the present merger, one has to wonder about antitrust scrutiny for any future
Google acquisitions – including vertical acquisitions that might otherwise appear benign in
traditional commercial settings. Questions about the prospect that Google might leverage a broader
search monopoly into dominance of distinct vertical search markets through acquisitions seem
destined to crop up again. Such questions will persist if Google’s suspected online search monopoly
is or is becoming entrenched, as will questions about how fairness and neutrality in something as
complex and subjective – and necessarily lacking in transparency – as search engine algorithms can
ever be monitored and effectively regulated.
From our perspective, governmental efforts to protect against manipulation of algorithms by
Google are not only likely to be ineffective, but they will necessarily raise First Amendment
questions as the government participates in decisions about the prioritization of information
reaching the public. Maintaining competitive markets for both general and niche search may be the
only alternative, ultimately, to an unregulatable monopoly. It is therefore appropriate for the
government to work within a public vision of longer-term developments and to place the present
acquisition within such a context.
INTRODUCTION
In this white paper, we present our analysis of the antitrust issues raised by a Google/ITA
combination. Among other things, we explore what Google’s dominant presence in the broader
online search market means for purposes of this transaction. We examine the possible nature of
Google’s presence in the market for back-end technology solutions that can access airfare and seat
availability information, which Google would be entering directly by acquiring ITA. We also
consider the relationship between the online travel search market and the back-end technology
solution market, and what that relationship suggests regarding this merger. Our analysis is based on
discussions with industry personnel, counsel, and a review of publicly available data and information.
We draw no conclusions about whether the acquisition violates the Clayton Act, which is likely to
depend on information that is not publicly available. In conducting its economic analysis, reviewing
2
the internal documents of the merging parties, and soliciting the views of market participants, the
DOJ will be tasked with answering these questions and many others.
BACKGROUND
OTAs, such as Expedia, Orbitz, and Travelocity, meta-search sites, such as Kayak and Bing Travel,1
and airlines and others that compete in the market for online travel search, rely on head-spinning
technology to provide Internet users with online access to accurate airfare pricing and seat
availability data. These firms require a back-end technology input that can agglomerate flight
schedules, routes, seat availability, tariffs, rules and prices, and then index that information and
convert it to a searchable format. Compounding matters, the technology must enable automated
searching in near-real time to account for query volume and accuracy, respectively.
OTAs, meta-search sites, and airline sites have a limited number of options in solving this difficult
problem, which is a prerequisite for their healthy existence. One option is to obtain the information
they need from one of three available global distribution system (GDS) suppliers, Travelport, Sabre
Holdings (Sabre) or Amadeus.2 Another option is to license a proprietary software program
developed by ITA.
ITA offers a range of products that service online travel search businesses and consumers, but the
company’s crown jewel is its QPX software solution. QPX can conduct automated searches of
airlines’ internal reservation systems, aggregate the relevant data, and provide it to OTAs and others
in a searchable format, while delivering all the necessary information to book and ticket an itinerary.
Since the company was founded in 1996, its QPX software has become very successful, attracting
reputable OTA and meta-search licensees such as Orbitz, Kayak, Hotwire and others, as well as
airlines like United, Continental, Southwest, US Airways and American Airways.
MARKET DEFINITION
Because of the range of players and products interwoven through the online travel search supply
chain, the effects of the transaction could conceivably reverberate throughout several product
markets. One of the DOJ’s tasks, as in any merger, will be to define antitrust product markets
affected by the transaction in an economically relevant way. Theoretically, the merger could impact
1 As a general matter, meta-search sites and OTAs differ primarily in the scope of their offerings. Like OTAs, meta-
search sites aggregate pricing and availability data for comparison shopping. But unlike OTAs, meta-search sites do not
directly offer seat booking. Further differences that might impact market definition are discussed in the next section.
2 GDSs collect and store information about airline pricing and seat availability, as well as related travel services like
hotels, rental cars and rail companies, creating an electronic marketplace to facilitate display and distribution of travel-
related offerings. Because GDSs do the above, they can act as a source of pricing and availability data for meta-search
sites. Because they can also access airlines’ internal reservation systems, they can go further for OTAs by enabling their
ability to book seats.
3
the markets for online search (where search engines compete for consumers), online search
advertising (where search engines compete for advertisers), online travel search (where OTAs, meta-
search sites and airline sites may all compete for consumers), online travel search advertising (where
the same may compete for advertisers), online travel search technology inputs (where ITA, GDSs
and others may all compete for the business of online travel search firms), and airline distribution of
pricing and availability data and airline ticket sales (where airlines compete to attract travelers). Most
immediately, the transaction seems significant for its potential competitive effects in online travel
search and online travel search advertising, assuming those markets are distinct from online search
and online search advertising, respectively. As we will discuss later, however, potential competitive
effects in some of the other markets referenced above also merit attention.
We suspect that online search and online travel search will be viewed as distinct markets. The
markets have in common that they generate two kinds of demand, consumer demand and advertiser
demand. But from a consumer perspective, online search and online travel search products likely
are not good substitutes. Online travel search is powered by a technology input that provides
consumers with access to airline pricing and availability data, whereas online search offers
consumers only traditional search-engine results.3 Whether advertisers might view online search
inventory and online travel search inventory as good substitutes is less clear, though it may be true
that the distinctions separating consumer demand for online search and online travel search
correspond with distinctions separating advertiser demand for each, which would suggest the
products are likely not good substitutes for advertisers either. Based on publicly available
information, we can only hypothesize that the two products’ respective utility to consumers,
underlying costs, and utility and pricing for advertisers differ to the point that a monopolist of one
would not lose so many sales to the other as to make a small but significant and non-transitory price
increase unprofitable. The DOJ’s internal analysis and the merging parties’ internal documents, as
well as the Division’s private conversations with market participants in online travel search, should
elucidate these questions further. If online search and online travel search are not good demand
substitutes, the product markets will be distinct for antitrust purposes.
Assuming online search and online travel search are distinct product markets, the transaction more
obviously threatens to alter the competitive status quo in online travel search and online travel
search advertising than in online search and online search advertising. To be sure, the transaction’s
potential effects in online search are an important aspect of the analysis, particularly to the degree
that the transaction threatens to raise barriers to entry in online search. But the consumer and
advertiser markets for online travel search face the prospect of significant changes in the nature of
competition in myriad ways. These issues, and related issues concerning how the transaction affects
the market for technology inputs into online travel search products, form the basis for much of the
3 As we will discuss below, we recognize that online search seems intimately connected to online travel search,
particularly insofar as online search impacts both supply and demand for online travel search products. Yet we also
recognize that traditional market definition analysis may not capture this connection. Resolving this dissonance is one of
the important analytical challenges posed by this transaction, as is often the case in mergers having aspects of systems
competition. See Gregory Gundlach & Diana L. Moss, Systems Competition and Challenges to Antitrust Thinking: An
Introduction, ANTITRUST BULL. (forthcoming 2011).
4
analysis that follows. We do, however, address the potential effects of the transaction in broader
online search. We also examine, near the end of this paper, whether and how the transaction might
impact the market for distribution of airline pricing and availability data, along with airline ticket
sales.
Even within online travel search, further relevant market distinctions are possible and should be
considered. OTAs and meta-search sites seem both to compete in online travel search, but the
former are built primarily to attract customers to book seats, while the latter are built primarily to
attract advertisers by giving consumers supposedly more robust comparison shopping options. Do
consumers and advertisers distinguish between these products in an economically relevant way? And
what about “search-engine travel search” products, like Microsoft’s Bing Travel, which operate
much like meta-search sites but are housed in a broader search engine environment? These
products can reach consumers both at early stages of travel planning, where users are exploring
travel ideas through the search engine, and at later stages of travel planning, where users are ready to
comparison shop airfares and book seats. Furthermore, where do airlines’ proprietary websites,
where users can search airline-specific pricing and availability data, fit into the market definition
equation? The DOJ should examine the prospect of narrower relevant product markets and how
any dispositive distinctions impact market shares and concentration, potentially heightening the risk
of anticompetitive effects.
One of the confounding aspects of this transaction is that, while the risk of anticompetitive effects
seems highly significant in online travel search and online travel search advertising, neither Google
nor ITA is directly in these markets in a traditional sense, nor is either company directly entering
these markets as a result of the transaction. This fact does not call for blindly dismissing the
potential for anticompetitive effects in these markets, but should instead prompt inquiry into
whether systems competition considerations should impact market definition analysis.4 Academic
marketing literature recognizes that businesses often compete for customers’ “pathways to
purchase,” whereby strategies are aimed at influencing customers on a broad spectrum as they move
through channels en route to purchasing a product. With respect to air travel, for example, an
online pathway to purchase may begin with a search-engine search exploring a destination’s
attractiveness, before proceeding to a meta-search cite to comparison shop airfares, before booking
a seat and completing a purchase on an OTA or airline website. It is appropriate for the DOJ to
consider whether various firms involved at different stages of online travel research and commerce
compete as parts of a system in which search engines, various types of airfare distributors, seat
booking engines, and airlines attempt to steer customers through a particularized pathway to
purchase. This view recognizes the impact that both Google and ITA might currently have in online
travel search and online travel search advertising, even though neither directly offers products in
those markets or will do so directly as a result of the transaction. Potential competition analysis can
play a role in recognizing this impact as well.
4 See supra note 4.
5
STRUCTURAL VIEWS OF THE TRANSACTION
Google’s proposed acquisition of ITA seems to have vertical, horizontal and conglomerate elements,
but not to fit squarely into any of these categories. Because Google is not directly in the market for
online travel search, and will not be directly in that market as an immediate result of acquiring ITA,
horizontal issues are not immediately evident. But the merger would enable and incent Google to
expand into online travel search in the future, using its ITA base to perhaps develop a meta-search
product, if not an OTA product. If this possibility is taken seriously, one could view the merger as
having a potential horizontal aspect.
At first blush the transaction seems to lend itself to vertical analysis because Google is acquiring an
input in the online search or online travel search supply chain. But this is problematic in that the
deal does not directly accomplish vertical integration in either of the two assumed product markets.
While the transaction otherwise might fit as a conglomerate merger, this would require a
determination that ITA’s software is a distinct product rather than an input into a product supply
chain, which seems wrong. For reasons discussed below, AAI believes the transaction is apt to be
viewed primarily from a vertical perspective, but it should be considered from a horizontal
perspective as well.
In one view, Google can be seen as taking steps to vertically integrate in online search by adding a
presence in the back-end technology solution market to its presence in the broader online search
market. Right now Google users can query Google’s search engine for flights between city pairs or
airport pairs (or a combination of the two) and return a Google flight search tool linking users to
any participating OTA. Such a query would also return airline advertisements linking users directly
to airline websites, along with OTA advertisements, other advertisements, and natural search results.
Post-merger, Google has hinted that it would endeavor to provide pricing and availability data
directly on its results page, though it would still direct users to other sites – presumably OTAs or the
airlines directly – to book seats. Google says it does not intend to enter the seat booking business.
Framing the transaction as a movement toward vertical integration in online search would cast ITA’s
software as a content-providing input into Google’s current online search product. However,
merely acquiring ITA does not directly integrate ITA’s content with Google’s existing search
product. Google must still develop an application for ITA’s content to its existing search product.
In another view, Google could be seen as vertically integrating in online travel search rather than
broader online search, but this too would be accomplished indirectly rather than directly as a result
of the transaction. Google first has to develop and introduce an online travel search product upon
closing the acquisition, as it has hinted it will. Upon doing so, Google apparently would become
one of three companies that are vertically integrated in online travel search for owning both a front-
end search product and a back-end technology solution. The others are Expedia, which is a
successful OTA and has a back-end product called Best Fare Search, and Sabre, which owns
Travelocity and has its own GDS. We are not aware of any other companies in online travel search
that are not dependent on outside partners for back-end technology solutions.
6
Framing the transaction as a movement toward vertical integration in online travel search would cast
Google’s future online travel search product as distinct from its existing, broader online search
product, with ITA’s software as a content-providing input into Google’s future online travel search
product. Again, however, Google must develop its new online travel search product before vertical
integration takes place.
The relevance of the distinction between characterizing ITA’s software as an input into Google’s
current online search product versus its future online travel search product may seem significant, but
market definition considerations counsel otherwise. Because online search and online travel search
seem likely to be distinct product markets, it matters little whether, technically speaking, Google will
be upgrading an existing product or creating a new product. What’s significant for antitrust
purposes is that Google is likely entering a new antitrust product market in either case, and in either
case it is acquiring an input for the new or changed product in advance of developing the new or
changed product itself. Therefore, while the transaction has vertical characteristics insofar as
Google is acquiring an input for either a new or changed product, it is not perfectly vertical because
a step remains between Google’s acquiring ITA and accomplishing vertical integration.
That the transaction is neither primarily horizontal nor perfectly vertical would suggest that it could
be a conglomerate merger of complementary products. However, ITA seems more like a content
provider than a manufacturer of a distinct product. Its QPX software is uniquely designed to
process airline data for purposes of displaying that data over the Internet, which serves relatively
narrow purposes of primary use to participants in the online travel search market. Furthermore,
Google’s public discussion of the transaction has been centered almost entirely on what ITA can do
for Google in online travel search. It has not hinted at any designs for wider-ranging applications
for the technology, which suggests Google views it more as an input than as a distinct product.
VERTICAL ANALYTICAL APPROACH AND THEORIES OF HARM
Under a traditional vertical analysis, the DOJ might focus on unilateral effects involving dominance
and foreclosure. Specifically, the Division might focus on whether Google would be acquiring
market power in the online travel search market or the back-end technology input market and
whether the transaction would have the effect of foreclosing existing or future competitors.
However, the Division can and should also inquire, more generally, whether the transaction creates
in the merged firm the ability and incentive to engage in exclusionary conduct directed at smaller,
travel-related rivals, or it may create or raise barriers to entry. This analytical approach is more in
keeping with modern thinking about vertical merger standards as it has evolved over the last two
decades.
LEVERAGING AND UPSTREAM MARKET POWER
7
A primary inquiry in standard vertical merger analysis is whether the transaction is likely to create or
enhance market power in affected markets, in this case the online travel search market and the back-
end technology input market. In the online travel search market, Google currently has at most an
attenuated presence, assuming that market is meaningfully distinct from the broader online search
market. 5 And as we have noted, acquiring ITA does not directly provide Google a competing
online travel search product. One could therefore take the view that the transaction does not pose
any risk of Google acquiring market power in online travel search; however, this is a myopic view
that ignores Google’s broader online search dominance.
Understanding whether the transaction is likely to create market power in Google in online travel
search requires careful examination of the risks under recognized monopoly leveraging theory.
While there is some doubt in the law as to whether there is anything untoward about leveraging a
monopoly in one market to achieve a competitive advantage in a second market,6 it is more clearly
unlawful for a firm to leverage monopoly power from one market into a second market where there
is a dangerous probability of achieving monopoly power in the second market.7
If Google has monopoly power in the broader online search market, which is widely suspected if not
yet confirmed by a court, the DOJ should determine whether the ITA acquisition may enable and
incent Google to leverage that power into monopolizing online travel search, or come dangerously
close, be it through vertical restraints like product bundling, exclusive dealing contracts, or other
opportunities that may come available post merger. To some degree a monopoly leveraging analysis
may be redundant, insofar as it is coextensive with foreclosure analysis, discussed below. But the
DOJ should still evaluate, even absent foreclosure, the danger that Google may monopolize online
travel search by way of any unique advantages owing to its online search dominance.
One question in this respect is whether Google’s search engine could by itself substantially
accommodate a dominant share of the online distribution needs of U.S. airlines, and the likelihood
that such an arrangement could come to pass even absent foreclosure in the online travel search
market. Part of determining that likelihood turns on whether ITA’s technology paired with
Google’s search engine would give Google a meaningful degree of control over both supply and
demand for travel information in the broader online search market, and whether control of supply
and demand in the broader online search market could afford the means to monopolize online travel
search. To the extent that ITA gives Google the ability to provide pricing and availability data in the
broader online search market, and its only competitor with respect to providing such data in the
broader online search market – Bing Travel – is an ITA licensee, Google arguably would control the
supply of travel pricing and availability information to consumers in the broader online search
5 See supra note 4.
6 See ABA SECTION OF ANTITRUST LAW, ANTITRUST LAW DEVELOPMENTS 286-89 (5th ed. 2002); but see Berkey Photo,
Inc. v. Eastman Kodak Co., 603 F.2d 263 (2d. Cir. 1979).
7 See, e.g., Eastmen Kodak Co. v. Image Technical Servs., Inc., 504 U.S. 541 (1992); accord Roger D. Blair & Amanda K.
Esquibel, Some Remarks on Monopoly Leveraging, 40 ANTITRUST BULL. 271 (1995) (discussing clear unlawfulness of
monopoly leveraging in a Section 2 context). Theoretically efficiencies might be in play in a Section 7 context. On the
feasibility of monopoly leveraging in spite of single monopoly profit theory, see Einer Elhauge, Tying, Bundled Discounts
and the Death of the Single Monopoly Profit Theory, 123 HARV. L. REV. 397 (2009).
8
market. To the extent that Google has the capacity to control the algorithms that dictate the search
engine’s natural search results and which advertisers appear in paid results, it controls the
economically valuable order in which consumers discover online travel search options. In this sense,
it arguably controls demand for online travel search products. Meaningful control of both supply
and demand for online travel search would seem to pose a dangerous risk of monopolization.
Of course, outright manipulation of search algorithms, if detectable, might constitute an antitrust
violation in its own right, and public awareness of the skewing of search results would seriously
threaten to undermine Google’s core product, which is a reliable search engine that users trust. 8 So
even if Google would have the ability to control both supply and demand for travel information in
online search, it is questionable whether it would have the incentive.
A determination would also remain as to whether such control would be meaningful to the point of
facilitating dominance in online travel search. Google’s role as a gatekeeper to online travel search
is a matter of some dispute. In the course of touting its AdWords program to travel businesses,
Google has in the past claimed that 74% of consumers use a search engine “when researching
and/or purchasing travel products or services online.”9 One report suggests that currently, search
engines “continue to be the primary way that Internet users navigate to key industry categories,” and
Google is responsible for driving over 30% of the upstream traffic received by firms in the travel
category.10 However, Google’s own analysis, using data from another source, suggests that Google
is responsible for only about 8% of the traffic received by the top ten travel websites.11 However,
Google’s 8% figure factors out what Google considers “lazy search,” where for example a user
enters the name of an OTA and then links immediately to it.12 Google believes it is not driving the
user to the OTA in these circumstances because the user is already intent on finding it.13 Of course,
users may also use Google’s search engine to research OTA options and then proceed to an OTA
site without linking through Google, in which case Google does seem to be driving traffic even
though this may not be captured statistically. Aided by its internal analysis and access to private
8 Presumably the need for users to find a meaningful relationship between a search query and the query’s results would
act as an inherent check against the threat of anticompetitive manipulation of search algorithms. However, presumably
too there is some margin at which manipulation can be employed without unduly threatening that meaningful
relationship. The scope of that margin is an important consideration in the course of examining Google’s ability to
engage in monopoly leveraging.
9 Google AdWords, Solutions for Travel, http://www.google.com/ads/metrics_trav.html (last updated Oct. 2006; last
visited Feb. 14, 2011).
10 Experian Hitwise, Experian Hitwise Reports Google Monthly Share of Searches at 72 Percent in September, Oct. 8, 2010,
http://www.hitwise.com/us/press-center/press-releases/google-monthly-share-of-searches-at-72-percent-i/.
According to the site, Google was responsible for 30.3% of traffic to online travel firms in the month of September
2010, Yahoo for 4.22% and Bing for 3.55%. Id.
11 Google, Facts About Google’s Acquisition of ITA Software; Claims vs. Facts, http://www.google.com/press/ita/facts.html
(last visited Feb. 14, 2011). Google’s figures, which are derived using data from Compete.com, suggest that ads from
various other websites and instances where users type a web address directly into their browsers are significantly bigger
drivers of traffic to OTAs than Google. Id. Using Travelocity, Priceline and Expedia as a sample, Google’s analysis
suggests that between 75% to 79% of traffic to these OTAs is attributable to a combination of “other” sites and users
typing directly into their browser. Id.
12 Id.
13 See id.
9
information, the DOJ should determine whether Google’s advantages in potentially controlling
supply and demand for travel information in the online search market are such that the transaction
poses a dangerous probability of Google monopolizing online travel search.
The transaction also impacts the upstream market for back-end technology inputs, which Google
would be affirmatively entering through ITA. The prospect of Google/ITA market power or
monopoly in the back-end technology solution market is something that publicly available
information does not clarify. It seems clear that ITA’s software system is a superior product, which
likely precipitated Google’s interest. ITA can claim Orbitz, Kayak, Hotwire, Bing Travel and other
OTAs and meta-search sites as licensees, as well as a host of airlines and others. Furthermore,
Hotwire and TripAdvisor, which are owned by Expedia and presumably could opt for Expedia’s
own Best Fare Search, use ITA instead.14 And reports have estimated that ITA currently powers a
majority of all online flight searches. However, Priceline, Travelocity and Expedia – three leaders in
online travel search – do not use ITA. Many successful airlines do not use ITA either. And the
existence or significance of barriers to entry in the back-end technology solutions market is not
immediately evident.
In inquiring whether the transaction creates market power for the merged firm in back-end
technology inputs, the DOJ will have to define the contours of this market and identify ITA’s
competitors, market share, and product substitutes. Even if the Division determines that ITA has
market or monopoly power in back-end technology inputs, however, this would not necessarily
prevent Google from consummating the transaction. The pertinent inquiry is whether anything
merger-specific may create or enhance market power. Google’s broader online search dominance is
again implicated in this inquiry; the Division should determine whether Google’s unique
characteristics as a future supplier of ITA’s product may enhance any market power ITA may
currently enjoy. The DOJ is likely to make a thorough analysis of this market, aided by the merging
companies’ internal documents and discussions with other market participants. The results of this
analysis will be critical in determining the transaction’s potential for exclusionary effects.
FORECLOSURE
Traditional Foreclosure of Existing or Potential Entrants
14That two wholly owned subsidiaries of Expedia would forego apparent cost advantages from using Expedia’s own
back-end solution raises interesting questions that warrant investigation. On the one hand, if the companies are
foregoing a marginal cost advantage to use ITA’s product, this could potentially speak to high switching costs or QPX’s
uniqueness and superiority as an input. On the other hand, it bears noting that Expedia’s Best Fare Search solution is
approximately ten years old, that Expedia acquired Hotwire and TripAdvisor after Best Fare Search was already
developed, and that Expedia’s Chief Financial Officer has suggested that Best Fare Search was built as an “in-house
solution” that is not readily adaptable to being commercialized for third-party use. See TNooz.com, Google-ITA Software
Deal: Expedia Boss Issues Stern Warning, July 30, 2010, http://www.tnooz.com/tag/best-fare-search/. It may simply be the
case that Hotwire and TripAdvisor use ITA because it is more affordable than adapting Best Fare Search to fit their
needs.
10
The DOJ should also examine whether the transaction likely may have significant foreclosure
effects, preventing either existing or future market entrants from competing. Post-merger, it can be
argued that Google will have the capacity to foreclose online travel search competition directly or
indirectly. Directly, Google could refuse to license the ITA software to OTAs and others that are
dependent on the software. Indirectly, Google could deter entrants by licensing the software on
selective or restrictive terms (e.g. devoid of technological upgrades and improvements). A key
inquiry in foreclosure analysis involves determining whether ITA’s input and the inputs of GDSs
and others are adequate substitutes, and in turn whether ITA’s input is an essential input. If the
inputs are not substitutes, then ITA’s input could be deemed essential. If that is the case,
foreclosure effects from the transaction would be too likely to ignore.
But Google has suggested that alternatives are available, in the form of GDSs, Expedia’s Best Fare
Search solution (although Expedia apparently has not yet commercialized Best Fare Search for third-
party use15), and others. Other leading OTAs apart from Expedia, including Priceline and
Travelocity, already use alternative solutions. Furthermore, it should be noted that ITA does not
control the underlying data itself, but rather a proprietary means of aggregating and processing that
data. The transaction therefore would not prevent others from entering and innovating in the
market for back-end airline pricing and availability solutions, though it remains for the DOJ to
determine whether barriers to entry are too significant. Given the complexity of ITA’s technology
and the difficulty of the problem it solves, barriers to entry could prove to be high. Another
question, which publicly available information does not clarify, is whether ITA’s technology has any
lock-in effect, and whether switching costs are high enough that ITA could raise the price of its
software a small but significant and non-transitory amount without losing market share.
Ability and Incentive to Engage in Exclusionary Conduct
Modern thinking about vertical merger standards recognizes that traditional foreclosure theory is
only one of several means by which a vertical merger can have exclusionary effects. The critical
inquiry is whether the transaction creates in a combined Google/ITA the ability and incentive to
engage in exclusionary conduct directed at smaller, travel-related rivals, including raising rivals’ costs
and creating barriers to entry or expansion. Such conduct can offend Section 7 of the Clayton Act.
Raising Rivals’ Costs
Traditional foreclosure analysis does not adequately capture the means through which a vertical
merger can affect competition in upstream and downstream markets. Post-merger, and upon a
sustainable prediction that Google will enter online travel search with its own product, Google will
be vertically integrated and competing primarily with un-integrated firms, downstream in the online
travel search market and upstream in the back-end technology input market. With the exception of
Expedia and Sabre’s Travelocity, we are aware of no other online travel search firms that do not
have to rely on outside partners for back-end technology inputs. The DOJ should ask: will a
15See TNooz.com, Google-ITA Software Deal: Expedia Boss Issues Stern Warning, July 30, 2010,
http://www.tnooz.com/tag/best-fare-search/.
11
combined Google/ITA lead to an increase in the price of back-end technology solutions for OTAs,
meta-search sites and airlines, thereby creating and increasing market power for Google? Will the
combination lead to lower revenues and in turn higher marginal costs for GDSs and other back-end
technology solutions, thereby creating and increasing market power for ITA’s QPX product?
Even if ITA’s input is non-essential and firms could switch if ITA raised the price of QPX by a
small but significant and non-transitory amount, the prospect of anticompetitive effects is not
rendered a nullity if the vertically integrated, merged firm would have market power upstream and
would be competing with un-integrated competitors downstream. Such a scenario creates
conditions that could allow the merged firm to “squeeze” rivals, and the DOJ should examine
whether the merged firm would have a post-merger ability and incentive to do so, giving rise to a
risk of exclusionary effects. For example, if Google’s downstream entry in online travel search will
likely lead to reduced demand and lower prices for downstream competitors’ advertising inventory,
the merged firm could squeeze downstream competitors by reducing, short of eliminating, access to
QPX if reduced access to QPX would lead to increased demand and higher prices for back-end
technology inputs. Likewise, the merged firm could squeeze upstream rivals (e.g. GDSs) by taking
the opposite tack of expanding access to QPX, thereby reducing demand for, and in turn revenue
from, competing upstream products, such that upstream alternatives become less of a competitive
constraint on QPX or are induced to exit the market. Google has said it intends to honor all
competitors’ existing contracts with ITA, and most online travel search firms that oppose the
transaction have expressed the opposite fear that Google would not renew such contracts or
otherwise refuse to do so on adequate terms. The DOJ should consider the prospect and effects of
both a contracted and an expanded presence for a Google-owned QPX in the back-end technology
input market.
At first blush, these potentially anticompetitive strategies might seem a good fit for a combined
Google/ITA. But they can only come to pass if Google successfully enters the downstream markets
for online travel search and online travel search advertising and achieves market power in the
upstream market for back-end technology solutions. As we have reiterated, Google currently does
not have an online travel search product to display pricing and availability data, and acquiring ITA
does not directly give it one. For that reason, the sustainability of a prediction that Google will
successfully enter online travel search directly, including by leveraging a monopoly power in broader
online search to achieve market power in online travel search, is critically important to evaluating the
transaction. The sustainability of a prediction that a merged Google/ITA would have market power
upstream in back-end technology inputs is equally important. The DOJ must determine whether
and how quickly Google is likely to become a significant player in online travel search, capable of
rapidly taking market share, and whether ITA’s product as supplied by Google would likely come to
dominate back-end technology inputs. The merging companies’ internal documents discussing
strategic motivations for the deal, which the DOJ can access but AAI cannot, will be crucial in this
and many other regards.
12
It bears noting that, in the wake of the Supreme Court’s decision in Pacific Bell v. Linkline, “price
squeeze” conduct like that described above does not amount to a Section 2 violation, though it may
be actionable if the conduct is linked to price predation or amounts to a violation of an antitrust
duty to deal.16 In a forward-looking Section 7 context, the DOJ should examine whether the merged
firm has incentives to engage in price predation or whether a price squeeze could amount to altering
an existing course of dealing with anticompetitive intent. One or both risks presumably would have
to be present before the Division might act upon price-squeeze concerns.
Given the Supreme Court’s opinion in Brooke Group Ltd. v. Brown & Williamson Tobacco Corp., the
DOJ should evaluate the prospect of price predation by examining the likelihood that the merged
firm would offer below-cost pricing with a dangerous probability of recouping its sacrifice.17
Google’s incentives to engage in price predation do not seem strong under this standard. First,
Google advertising inventory is sold using keyword auctions, a mechanism that would naturally tend
to prevent Google from offering inventory at artificially low prices. Even under a hypothetical
alternative pricing model, Google could likely achieve a price advantage without resorting to below-
cost pricing because of the marginal cost advantage it would enjoy as a vertically integrated firm
competing with un-integrated firms. Second, if Google were to undertake a foreclosure strategy
aimed at online travel search competitors, it would stand to lose an assured and probably substantial
income flow both from licensing ITA’s QPX product and selling Google advertising inventory to
these firms. The Division would have to determine that the losses sacrificed under such a strategy
would not be too great as compared to the possible benefits of a monopoly, discounting for the risk
that the strategy might fail.
However, whether a price squeeze strategy might be paired with breach of an antitrust duty to deal is
a different question. While antitrust law imposes no generalized duty to deal, district courts post-
Linkline have held that voluntarily altering a course of dealing that induced reliance by competitors,
when coupled with “anticompetitive malice,” can be actionable in antitrust.18 Here, it may be
plausible that online travel search firms could have been induced to license QPX by reliance on
ITA’s status as a stand-alone input supplier, motivated to broadly and continually license its product.
Thus the Division should consider whether, through the transaction, ITA would be voluntarily
altering its status so as to give rise to an antitrust duty to deal. If ITA’s voluntary decision to fold its
current business into the business of a firm that will compete with its current customers creates a
duty in ITA to continue its role in providing its current customers with their current capabilities, and
yet the merged firm would rationally pursue an anticompetitive price squeeze, a cognizable theory of
harm may arise under Section 7.
Creating Barriers to Entry or Expansion
16 129 S. Ct. 1109 (2009).
17 509 U.S. 209 (1993).
18 See Safeway Inc. v. Abbott Laboratories, 2010 WL 147988 (N.D. Cal. Jan. 12, 2010) (citing Aspen Skiing Co. v. Aspen
Highlands Skiing Corp., 472 U.S. 585 (1985) and Verizon Communications Inc. v. Law Office of Curtis V. Trinko, LLP,
540 U.S. 298 (2004)).
13
The DOJ should also inquire whether the merger creates or raises barriers to entry or expansion that
would lead to anticompetitive harm. Vertical integration can itself be a barrier to entry where
entrants recognize the difficulty in competing with a large firm like Google after it adds a marginal
cost advantage for owning a back-end technology input supplier to its already tremendous scale and
network advantages. However, barriers to entry would pose an even greater concern if, post-
merger, new or existing entrants could adequately compete only by entering the market at two levels,
i.e. in both upstream back-end technology inputs and downstream online travel search products.
Post-merger, online travel search entrants seeking to use ITA as a back-end technology input might
be deterred because their access to a critical input would be controlled by a potential competitor,
which poses an obvious business risk. They might also be deterred in developing or enhancing a
product tied to ITA if doing so would require disclosure of competitively sensitive information, such
as technological innovations, to a potential competitor in Google. It seems clear that incorporating
a back-end technology input with a front-end travel search product invariably requires information
exchange, but publicly available information does not clarify whether such exchange necessarily
demands disclosure of competitively sensitive information. If the DOJ’s investigation reveals that
these concerns are significant, and existing alternatives to ITA, such as GDSs, are not easily available
to entrants or are meaningfully inferior to ITA, the risk that the transaction might create
competitively dangerous barriers to entry seems strong.
The DOJ should examine whether GDSs and/or alternative back-end solutions could adequately
accommodate online travel search competitors’ needs in a manner that would mitigate the risk of
barriers to entry being created by Google controlling ITA. If GDSs and/or alternatives could
substantially accommodate online travel search competitors, or if the risks of disclosing
competitively sensitive information to a future competitor through post-merger licensing of QPX
are not serious, the merger may not raise barriers to entry in a manner that offsets proffered
efficiencies.19 However, if the Division finds the opposite is true, and new or existing entrants
would have to enter both the back-end technology input market and the online travel search market
in order to effectively compete, the Division should definitively resolve this risk. A need to enter at
two levels might lead to consolidation at both levels, with existing OTAs and GDSs merging as a
competitive response to Google/ITA. In addition, new entrants in online travel search might
effectively be foreclosed for having to overcome barriers to entry in the input market alone. As
noted previously, publicly available information does not clarify the significance of entry barriers in
the input market, but the complexity of the problem that ITA and GDSs solve and the advanced
technology that these firms deploy may suggest that entry barriers are high.
Finally, the DOJ should consider whether the merger raises barriers to entry in the broader online
search market, where Google faces any competition for its dominant market share primarily from
19
A
substantial
body
of
literature
addresses
efficiencies
associated
with
vertical
mergers.
See,
e.g.,
M.
Howard
Morse,
Vertical
Mergers:
Recent
Learning,
53
BUS.
LAW.
1217,
1223
(1998)
(“(i)
the
reduction
of
transaction
costs,
where
the
elimination
of
negotiation
and
execution
of
contracts
reduces
associated
risk,
uncertainty,
and
opportunism;
(ii)
the
heightened
coordination
of
or
synergistic
improvements
in
design,
production,
and
distribution;
and
(iii)
the
elimination
of
double
marginalization
through
internalization
of
pricing
decisions.”)
14
Bing and Yahoo. While in the abstract there is nothing competitively suspicious about a search
engine seeking to expand into a vertical search market that is part of the largest existing e-commerce
category (valued at $80 billion in 2009), the circumstances here are that online travel is one of the
few areas where Google’s closest legitimate challenger in broader online search, Microsoft, has made
competitive inroads, and Google has chosen to enter online travel search by acquiring a technology
input upon which Microsoft relies. Because of Google’s dominance in broader online search, the
Division should be especially wary of allowing an acquisition that would have the effect of
preventing competitors in broader online search from maintaining a level of profitability through
vertical markets that might be integral to fostering sustainable entry in broader online search.
HORIZONTAL ANALYSIS
Potential Competition
Although the transaction more obviously lends itself to vertical considerations, horizontal issues do
arise and merit attention. Specifically, the transaction should be evaluated for its effects on potential
competition in online travel search. Potential competition is premised on the idea that under certain
circumstances, a firm positioned on the edge of a market can exert a competitive presence within the
market even if it is not actually competing there. If, absent the merger, Google and/or ITA could
nonetheless be considered a potential entrant into online travel search, a merger eliminating either’s
status as such should be evaluated for its effects in online travel search, much as a merger
eliminating an incumbent in online travel search would be evaluated similarly. Likewise, if the
merged firm would become a potential entrant into online travel search as a result of the
combination, the merger should be evaluated for the effects that a combined Google/ITA would
have on competition in online travel search.
The DOJ has recognized harm to both actual and perceived potential competition, the former where
a merger impacts a potential entrant’s prospect of entry through more pro-competitive means, and
the latter where a merger impacts the constraining influence that a potential entrant may have on the
behavior of incumbent firms. Among other things, both theories presume that a firm will only be a
potential entrant to a market if it is positioned at the edge of that market and if it is the only or one
of the very few potential entrants into a highly concentrated market.
Commentators have recently argued, and we agree, that distinctions between actual and perceived
potential competition are largely irrelevant and unhelpful, because “Courts will hold that a merger
violates Section 7 only if each link in the causal chain from the merger itself to the alleged
anticompetitive effect is sufficiently strong that the feared effect can be predicted with ‘reasonable
probability.’”20 This is true “[w]hether the relationship between the merging firms is described as
20Gregory J. Werden & Kristen C. Limarzi, Forward-Looking Merger Analysis and the Superfluous Potential Competition Doctrine,
77 Antitrust L.J. 141 (2010).
15
You can also read
